1. Field
One or more embodiments relate to speech signal transmission and reception apparatuses and speech signal transmission and reception methods, which compress and then transmit speech signals and restore received speech signals.
2. Description of the Related Art
Generally, a speech signal transmission apparatus transmits speech signals by splitting them into several parameters indicating characteristics of a speech source and a resonance system, based on the idea that the speech signals are regarded as an output of the resonance system excited according to the speech source, and a speech signal reception apparatus synthesizes original speech signals according to the parameters.
The speech signal transmission and reception apparatuses include codecs which encode and decode speech signals in a frame unit. Among such codecs, for example, a G.729 codec receives a frame from a frame part and encodes and decodes the speech signals in units of 10 ms.
The frame part classifies samples which are successively transmitted at 8 KHz from the exterior into samples of 10 ms and provides 80 classified samples as one frame to the G.729 codec as an input signal.
The G.729 codec may be achieved using a Digital Signal Processor (DSP).
In this case, a memory of the DSP includes a code part which generates and stores executable code corresponding to the number of processed channels and a data part, a program use space, which stores global variables, channel buffer stacks, and the like.
In this codec, the number of achievable channels is determined according to the processing capabilities of the DSP. If the number of channels capable of being processed in the DSP increases, since execution codes corresponding to the number of channels should be generated, the necessary amount of the memory also increases.
Furthermore, when loss compression data is needed during compression of a multichannel speech signal or when lossless data is needed to maximize performance, the amount of signals of speech data to be transmitted increases in correspondence to the number of microphones.
Moreover, when speech signals are collected through multichannel microphones, synchronization of speech signals varies according to the location or characteristic of the microphones and powers between the speech signals become different. Accordingly, compression is not easy and compression efficiency is low.